A Theory of Structure: Duality, Agency, and Transformation

A Theory of Structure: Duality, Agency, and Transformation Author(s): William H. Sewell, Jr. Source: American Journal of Sociology, Vol. 98, No. 1 (Jul., 1992), pp. 1-29 Published by: The University of Chicago Press Stable URL: . Accessed: 21/02/2014 16:35 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@. .

The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to American Journal of Sociology.



This content downloaded from 146.96.128.36 on Fri, 21 Feb 2014 16:35:54 PM All use subject to JSTOR Terms and Conditions

A Theory of Structure: Duality, Agency, and Transformation1

William H. Sewell, Jr. University of Chicago

"Structure"is one of the most important, elusive, and undertheorized concepts in the social sciences. Setting out from a critique and reformulation of Anthony Giddens's notion of the duality of structure and Pierre Bourdieu's notion of habitus, this article attempts to develop a theory of structure that restores human agency to social actors, builds the possibility of change into the concept of structure, and overcomes the divide between semiotic and material-

ist visions of structure.

"Structure"is one of the most important and most elusive terms in the vocabulary of current social science. The concept is central not only in such eponymous schools as structural functionalism, structuralism, and poststructuralism, but in virtually all tendencies of social scientific thought. But if social scientists find it impossible to do without the term "structure," we also find it nearly impossible to define it adequately. Many of us have surely had the experience of being asked by a "naive" student what we mean by structure, and then finding it embarrassingly difficult to define the term without using the word "structure" or one of its variants in its own definition. Sometimes we find what seems to

be an acceptable synonym-for example, "pattern"-but all such synonyms lack the original's rhetorical force. When it comes to indicating

' This article has benefited, during its many revisions, from the careful reading and constructivecriticismof a large number of friends and colleagues. Although I have sometimesfailed to heed their good advice, I am deeply gratefulto ElizabethAnderson, Jeffrey Alexander, Ronald Aminzade, Renee Anspach, Terry Boswell, Peggy Evans, Neil Fligstein, Steven Gudeman, Ronald Herring, Ronald Inden, David Laitin, BarbaraLaslett, Michael Kennedy, SherryOrtner, Sylvia Pedraza,Joan Scott, Ellen Sewell, Theda Skocpol, Ann Swidler, John Urry, Loic Wacquant, several reviewers, and the audiencesof seminarsand colloquiaat the Universityof Minnesota, the University of Michigan, HarvardUniversity, NorthwesternUniversity, the University of Chicago, and the Center for Advanced Study in the Behavioral Sciences (CASBS). This article was revised extensively while I was a fellow at the CASBS. I am grateful for support provided by the National Science Foundation, grant BNS870064,and by a fellowshipfromtheJohn SimonGuggenheimMemorialFoundation.

? 1992 by The University of Chicago. All rights reserved.

0002-9602/93/9801-0001$01.50

AJS Volume 98 Number 1 (July 1992): 1-29

1

This content downloaded from 146.96.128.36 on Fri, 21 Feb 2014 16:35:54 PM All use subject to JSTOR Terms and Conditions

American Journal of Sociology

that a relation is powerful or important it is certainly more convincing to designate it as "structural" than as "patterning."

The term structure empowers what it designates. Structure, in its nominative sense, always implies structure in its transitive verbal sense. Whatever aspect of social life we designate as structure is posited as "structuring"some other aspect of social existence-whether it is class that structures politics, gender that structures employment opportunities, rhetorical conventions that structure texts or utterances, or modes of production that structure social formations. Structure operates in social scientific discourse as a powerful metonymic device, identifying some part of a complex social reality as explaining the whole. It is a word to conjure with in the social sciences. In fact, structure is less a precise concept than a kind of founding or epistemic metaphor of social scientific-and scientific-discourse.2 For this reason, no formal definition can succeed in fixing the term's meaning: the metaphor of structure continues its essential if somewhat mysterious work in the constitution of social scientific knowledge despite theorists' definitional efforts.

There are, nevertheless, three problems in the current use of the term that make self-conscious theorizing about the meanings of structure seem worthwhile. The most fundamental problem is that structuralor structuralist arguments tend to assume a far too rigid causal determinism in social life. Those features of social existence denominated as structures tend to be reified and treated as primary, hard, and immutable, like the girders of a building, while the events or social processes they structure tend to be seen as secondary and superficial, like the outer "skin" of a skyscraper, or as mutable within "hard" structural constraints, like the layout of offices on floors defined by a skeleton of girders. What tends to get lost in the language of structure is the efficacy of human action-or "agency," to use the currently favored term. Structures tend to appear in social scientific discourse as impervious to human agency, to exist apart from, but nevertheless to determine the essential shape of, the strivings and motivated transactions that constitute the experienced surface of social life. A social science trapped in an unexamined metaphor of structure tends to reduce actors to cleverly programmed automatons. A second and closely related problem with the notion of structure is that it makes dealing with change awkward. The metaphor of structure implies stability. For this reason, structural language lends itself readily to explanations of how social life is shaped into consistent patterns, but not to

2 The term "structure" seems to play an essentially identical role in the natural sciences. Such usages originated, as far as I am aware, in 17th- and 18th-century botany, from which they spread to other natural and social sciences (see Foucault 1973, pp. 132-38).

2

This content downloaded from 146.96.128.36 on Fri, 21 Feb 2014 16:35:54 PM All use subject to JSTOR Terms and Conditions

The Theory of Structure

explanations of how these patterns change over time. In structural discourse, change is commonly located outside of structures, either in a telos of history, in notions of breakdown, or in influences exogenous to the system in question. Consequently, moving from questions of stability to questions of change tends to involve awkward epistemological shifts.

The third problem is of a rather different order: the term structure is used in apparently contradictory senses in different social scientific discourses, particularly in sociology and anthropology. Sociologists typically contrast "structure"to "culture." Structure, in normal sociological usage, is thought of as "hard" or "material" and therefore as primary and determining, whereas culture is regarded as "soft" or "mental" and therefore as secondary or derived. By contrast, semiotically inclined social scientists, most particularly anthropologists, regard culture as the preeminent site of structure. In typical anthropological usage, the term structure is assumed to refer to the realm of culture, except when it is modified by the adjective "social." As a consequence, social scientists as different in outlook as Theda Skocpol and Marshall Sahlins can be designated as "structuralists"by their respective disciplines. Sociologists and anthropologists, in short, tend to visualize the nature and location of structure in sharply discrepant, indeed mutually incompatible, ways.3

In view of all these problems with the notion of structure, it is tempting to conclude that the term should simply be discarded. But this, I think, is impossible: structure is so rhetorically powerful and pervasive a term that any attempt to legislate its abolition would be futile. Moreover, the notion of structure does denominate, however problematically, something very important about social relations: the tendency of patterns of relations to be reproduced, even when actors engaging in the relations are unaware of the patterns or do not desire their reproduction. In my opinion, the notion of structure neither could nor should be banished from the discourse of social science. But it does need extensive rethinking. This article will attempt to develop a theory of structure that overcomes the three cardinal weaknesses of the concept as it is normally employed in social science. The theory will attempt (1) to recognize the agency of social actors, (2) to build the possibility of change into the concept of structure, and (3) to overcome the divide between semiotic and material-

3 This bifurcationof the meaning of structureespeciallyinhibits communicationbetween two groupsof social scientistswhose currentprojectsseem convergentbut who have thus far paid little attention to one another:the growing band of sociologists who are examiningthe culturaldimensionsof social life and the anthropologistswho are insistingon the importanceof power and practicein understandingculture. For an assessmentof the mushroomingfield of culturalsociology, see Lamontand Wuthnow (1990).For trendsin currentanthropology,see the remarksof Ortner(1984, pp. 144-60).

3

This content downloaded from 146.96.128.36 on Fri, 21 Feb 2014 16:35:54 PM All use subject to JSTOR Terms and Conditions

American Journal of Sociology

ist visions of structure. My strategy will be to begin from what I regard as the most promising existing formulations-Anthony Giddens's notion of "the duality of structure" and, at a later point in the argument, Pierre Bourdieu's concept of habitus-and to develop a more adequate theory by means of critique, reformulation, and elaboration.4

THE DUALITYOF STRUCTURE:A CRITIQUEAND REFORMULATIONOF GIDDENS'STHEORY

The most sustained effort at reconceptualizing structure in recent social theory has been made by Anthony Giddens, who has been insisting since the mid-1970s that structures must be regarded as "dual"(Giddens 1976, 1979, 1981, 1984). By this he means that they are "both the medium and the outcome of the practices which constitute social systems" (Giddens 1981, p. 27). Structures shape people's practices, but it is also people's practices that constitute (and reproduce) structures. In this view of things, human agency and structure, far from being opposed, in fact presuppose each other. Structures are enacted by what Giddens calls "knowledgeable" human agents (i.e., people who know what they are doing and how to do it), and agents act by putting into practice their necessarily structured knowledge. Hence, "structures must not be conceptualized as simply placing constraints on human agency, but as enabling" (Giddens 1976, p. 161). This conception of human agents as "knowledgeable" and "enabled" implies that those agents are capable of putting their structurally formed capacities to work in creative or innovative ways. And, if enough people or even a few people who are powerful enough act in innovative ways, their action may have the consequence of transforming the very structures that gave them the capacity to act. Dual structures therefore are potentially mutable. It is no accident that Giddens calls his theory "the theory of structuration," indicating by this neologism that "structure" must be regarded as a process, not as a steady state.

As a theoretically self-conscious social historian, I find Giddens's notion of the duality of structure particularly congenial. Much of the best social history of the past quarter-centuryhas adopted an implicit theoretical strategy quite consistent with Giddens's theory. Social historians have

4 It is not my purposeto developa full critiqueor appreciationof GiddensorBourdieu. The critical literatureon both is growing rapidly. Held and Thompson (1989) and Bryant and Jary (1991)include not only a wide rangeof critiquesof Giddens'swork by prominentscholars but also useful bibliographicallistings of previous critiques. On Bourdieu,see DiMaggio(1979),Brubaker(1985),Lamontand Lareau(1988),and Wacquant(1989). The last of these containsextensive referencesto criticalworks on Bourdieu.

4

This content downloaded from 146.96.128.36 on Fri, 21 Feb 2014 16:35:54 PM All use subject to JSTOR Terms and Conditions

 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download